Influence of the Bounds of the Hyperparameters on the Reconstruction of Hubble Constant with Gaussian Process

TL;DR

This paper examines how the bounds of hyperparameters in Gaussian process models affect the accuracy and robustness of reconstructing the Hubble constant from observational data, emphasizing the importance of hyperparameter bounds.

Contribution

It investigates the impact of hyperparameter bounds in Gaussian process reconstruction of Hubble constant and highlights the necessity of considering these bounds for reliable results.

Findings

Hyperparameter bounds significantly influence H0 reconstruction.

Considering hyperparameter bounds improves the robustness of GP extrapolations.

Forecasts show bounds are essential for future data analysis.

Abstract

The cosmological model-independent method Gaussian process (GP) has been widely used in the reconstruction of Hubble constant $H_0$, and the hyperparameters inside GP influence the reconstructed result derived from GP. Different hyperparameters inside GP are used in the constraint of $H_0$ derived from GP with observational Hubble parameter $H(z)$ data (OHD), and the influence of the hyperparameters inside GP on the reconstruction of $H_0$ with GP is discussed. The discussion about the hyperparameters inside GP and the forecasts for future data show that the consideration of the lower and upper bounds on the GP's hyperparameters are necessary in order to get an extrapolated result of $H_0$ from GP reliably and robustly.